Merge branch 'main' into isearch-paste

Author: tanloong

Doc/c-api/perfmaps.rst

@@ -31,7 +31,7 @@ Note that holding an :term:`attached thread state` is not required for these API
    or ``-2`` on failure to create a lock. Check ``errno`` for more information
    about the cause of a failure.
 
-.. c:function:: int PyUnstable_WritePerfMapEntry(const void *code_addr, unsigned int code_size, const char *entry_name)
+.. c:function:: int PyUnstable_WritePerfMapEntry(const void *code_addr, size_t code_size, const char *entry_name)
 
    Write one single entry to the ``/tmp/perf-$pid.map`` file. This function is
    thread safe. Here is what an example entry looks like::

Include/cpython/ceval.h

@@ -38,7 +38,7 @@ typedef struct {
 PyAPI_FUNC(int) PyUnstable_PerfMapState_Init(void);
 PyAPI_FUNC(int) PyUnstable_WritePerfMapEntry(
     const void *code_addr,
-    unsigned int code_size,
+    size_t code_size,
     const char *entry_name);
 PyAPI_FUNC(void) PyUnstable_PerfMapState_Fini(void);
 PyAPI_FUNC(int) PyUnstable_CopyPerfMapFile(const char* parent_filename);

Include/internal/pycore_ceval.h

@@ -94,7 +94,7 @@ typedef struct {
     void* (*init_state)(void);
     // Callback to register every trampoline being created
     void (*write_state)(void* state, const void *code_addr,
-                        unsigned int code_size, PyCodeObject* code);
+                        size_t code_size, PyCodeObject* code);
     // Callback to free the trampoline state
     int (*free_state)(void* state);
 } _PyPerf_Callbacks;
@@ -108,6 +108,10 @@ extern PyStatus _PyPerfTrampoline_AfterFork_Child(void);
 #ifdef PY_HAVE_PERF_TRAMPOLINE
 extern _PyPerf_Callbacks _Py_perfmap_callbacks;
 extern _PyPerf_Callbacks _Py_perfmap_jit_callbacks;
+extern void _PyPerfJit_WriteNamedCode(const void *code_addr,
+                                      size_t code_size,
+                                      const char *entry,
+                                      const char *filename);
 #endif
 
 static inline PyObject*

Include/internal/pycore_debug_offsets.h

@@ -215,6 +215,7 @@ typedef struct _Py_DebugOffsets {
         uint64_t state;
         uint64_t length;
         uint64_t asciiobject_size;
+        uint64_t compactunicodeobject_size;
     } unicode_object;
 
     // GC runtime state offset;
@@ -370,6 +371,7 @@ typedef struct _Py_DebugOffsets {
         .state = offsetof(PyUnicodeObject, _base._base.state), \
         .length = offsetof(PyUnicodeObject, _base._base.length), \
         .asciiobject_size = sizeof(PyASCIIObject), \
+        .compactunicodeobject_size = sizeof(PyCompactUnicodeObject), \
     }, \
     .gc = { \
         .size = sizeof(struct _gc_runtime_state), \

Include/internal/pycore_interp_structs.h

@@ -69,7 +69,7 @@ struct code_arena_st;
 struct trampoline_api_st {
     void* (*init_state)(void);
     void (*write_state)(void* state, const void *code_addr,
-                        unsigned int code_size, PyCodeObject* code);
+                        size_t code_size, PyCodeObject* code);
     int (*free_state)(void* state);
     void *state;
     Py_ssize_t code_padding;

Include/internal/pycore_jit.h

@@ -23,7 +23,7 @@ typedef _Py_CODEUNIT *(*jit_func)(
     _PyStackRef _tos_cache0, _PyStackRef _tos_cache1, _PyStackRef _tos_cache2
 );
 
-_Py_CODEUNIT *_PyJIT(
+_Py_CODEUNIT *_PyJIT_Entry(
     _PyExecutorObject *executor, _PyInterpreterFrame *frame,
     _PyStackRef *stack_pointer, PyThreadState *tstate
 );

Include/internal/pycore_jit_unwind.h

@@ -0,0 +1,68 @@
+#ifndef Py_INTERNAL_JIT_UNWIND_H
+#define Py_INTERNAL_JIT_UNWIND_H
+
+#ifndef Py_BUILD_CORE
+#  error "this header requires Py_BUILD_CORE define"
+#endif
+
+#include <stddef.h>
+#include <stdint.h>
+
+#if defined(_Py_JIT) && defined(__linux__) && defined(__ELF__)
+#  define PY_HAVE_JIT_GDB_UNWIND
+#endif
+
+#if defined(PY_HAVE_PERF_TRAMPOLINE) || defined(PY_HAVE_JIT_GDB_UNWIND)
+
+#if defined(PY_HAVE_JIT_GDB_UNWIND)
+extern PyMutex _Py_jit_debug_mutex;
+#endif
+
+/* DWARF exception-handling pointer encodings shared by JIT unwind users. */
+enum {
+    DWRF_EH_PE_absptr = 0x00,
+    DWRF_EH_PE_omit = 0xff,
+
+    /* Data type encodings */
+    DWRF_EH_PE_uleb128 = 0x01,
+    DWRF_EH_PE_udata2 = 0x02,
+    DWRF_EH_PE_udata4 = 0x03,
+    DWRF_EH_PE_udata8 = 0x04,
+    DWRF_EH_PE_sleb128 = 0x09,
+    DWRF_EH_PE_sdata2 = 0x0a,
+    DWRF_EH_PE_sdata4 = 0x0b,
+    DWRF_EH_PE_sdata8 = 0x0c,
+    DWRF_EH_PE_signed = 0x08,
+
+    /* Reference type encodings */
+    DWRF_EH_PE_pcrel = 0x10,
+    DWRF_EH_PE_textrel = 0x20,
+    DWRF_EH_PE_datarel = 0x30,
+    DWRF_EH_PE_funcrel = 0x40,
+    DWRF_EH_PE_aligned = 0x50,
+    DWRF_EH_PE_indirect = 0x80
+};
+
+/* Return the size of the generated .eh_frame data for the given encoding. */
+size_t _PyJitUnwind_EhFrameSize(int absolute_addr);
+
+/*
+ * Build DWARF .eh_frame data for JIT code; returns size written or 0 on error.
+ * absolute_addr selects the FDE address encoding:
+ * - 0: PC-relative offsets (perf jitdump synthesized DSO).
+ * - nonzero: absolute addresses (GDB JIT in-memory ELF).
+ */
+size_t _PyJitUnwind_BuildEhFrame(uint8_t *buffer, size_t buffer_size,
+                                 const void *code_addr, size_t code_size,
+                                 int absolute_addr);
+
+void *_PyJitUnwind_GdbRegisterCode(const void *code_addr,
+                                  size_t code_size,
+                                  const char *entry,
+                                  const char *filename);
+
+void _PyJitUnwind_GdbUnregisterCode(void *handle);
+
+#endif  // defined(PY_HAVE_PERF_TRAMPOLINE) || defined(PY_HAVE_JIT_GDB_UNWIND)
+
+#endif  // Py_INTERNAL_JIT_UNWIND_H

Include/internal/pycore_optimizer.h

@@ -198,6 +198,7 @@ typedef struct _PyExecutorObject {
     uint32_t code_size;
     size_t jit_size;
     void *jit_code;
+    void *jit_gdb_handle;
     _PyExitData exits[1];
 } _PyExecutorObject;
 

Lib/pickle.py

@@ -920,17 +920,11 @@ def save_picklebuffer(self, obj):
                     # Write data in-band
                     # XXX The C implementation avoids a copy here
                     buf = m.tobytes()
-                    in_memo = id(buf) in self.memo
                     if m.readonly:
-                        if in_memo:
-                            self._save_bytes_no_memo(buf)
-                        else:
-                            self.save_bytes(buf)
+                        self._save_bytes_no_memo(buf)
                     else:
-                        if in_memo:
-                            self._save_bytearray_no_memo(buf)
-                        else:
-                            self.save_bytearray(buf)
+                        self._save_bytearray_no_memo(buf)
+                    self.memoize(obj)
                 else:
                     # Write data out-of-band
                     self.write(NEXT_BUFFER)

Lib/profiling/sampling/sample.py

@@ -58,6 +58,10 @@ def __init__(self, pid, sample_interval_usec, all_threads, *, mode=PROFILING_MOD
         try:
             self.unwinder = self._new_unwinder(native, gc, opcodes, skip_non_matching_threads)
         except RuntimeError as err:
+            if os.name == "nt" and sys.executable.endswith("python.exe"):
+                raise SystemExit(
+                    "Running profiling.sampling from virtualenv on Windows platform is not supported"
+                ) from err
             raise SystemExit(err) from err
         # Track sample intervals and total sample count
         self.sample_intervals = deque(maxlen=100)